AI outperforms clinicians’ judgment in triaging postop patients to ICU

Editor's Note

Artificial intelligence (AI) in the form of a machine-learned algorithm correctly triaged the vast majority of postoperative patients to the ICU, in this pilot study presented October 29 at the American College of Surgeons Clinical Congress 2019 in San Francisco.

The algorithm included 87 clinical variables and 15 specific criteria related to the appropriateness of admission to the ICU within 48 hours of surgery. Admission to the ICU was considered appropriate if one of these criteria was met.

Clinicians answered a questionnaire as to how they would evaluate the need for the ICU for each patient. The researchers asked the machine the same questions.

Results were compared:

• AI correctly triaged 41 of the 50 patients in the study (82%).
• Surgeons correctly triaged 35 patients (70%).
• Intensivists correctly triaged 32 patients (64%).
• Anesthesiologists correctly triaged 29 patients (58%).

The number of incorrect triage decisions was lowest for AI at 18%, followed by surgeons at 30%, intensivists at 36%, and anesthesiologists at 42%.

Though the algorithm clearly outperformed clinician's judgment, it is a first step, the researchers say. The researchers plan to apply the algorithm to other patient populations and include other demographic and clinical features.